Abstract: A noise robustness method operates jointly in a signal domain and a model domain. For example, energy is added in the signal domain for frequency bands where an actual noise level of an incoming signal is lower than a noise level used to train models, thus obtaining a compensated signal. Also, energy is added in the model domain for frequency bands where noise level of the incoming signal or the compensated signal is higher than the noise level used to train the models. Moreover, energy is never removed, thereby avoiding problems of higher sensitivity of energy removal to estimation errors.

Abstract: A method and system for differentiating between intended user input and inadvertent or incidental contact with a touchpad is herein disclosed. When a user engages the touchpad, sensors on the touchpad are activated and generate touch sensor signals. Based on the pattern of engaged sensors, a hand pattern can be determined. From the hand pattern, a hand model may be retrieved. The hand model may indicate passive zones and active zones. Contact in the active zones may be considered intentional, while contact in the passive zones may be considered unintended or incidental. Moreover, a global shift may be calculated, and input from the active zones may be compensated for the global shift. The input from the active zones can then be used to control a graphical user interface.

Abstract: System and method for touch sensitive surface provide detection and tracking of multiple touch points on the surface by using two independent arrays of orthogonal linear capacitive sensors.

Abstract: Sensors around the periphery of the remote control unit detect contact with the user's hand. A trained model-based pattern classification system analyzes the periphery sensor data and makes a probabilistic prediction of the user's hand size. The hand size is then used to control a mapping system that defines how gestures by the user's thumb upon a touchpad of the remote control unit are mapped to the control region upon a separate display screen.

Abstract: A graphic user interface system for use with a content based retrieval system includes an active display having display areas. For example, the display areas include a main area providing an overview of database contents by displaying representative samples of the database contents. The display areas also include one or more query areas into which one or more of the representative samples can be moved from the main area by a user employing gesture based interaction. A query formulation module employs the one or more representative samples moved into the query area to provide feedback to the content based retrieval system.

Abstract: The handheld case of the remote control unit includes at least one touchpad, and other sensors, such as acceleration sensors, case perimeter sensors, pressure sensors, RF signal sensors. These sensors provide a rich array of sensory inputs that are classified by a pattern recognizer to generate control commands for both the consumer electronic equipment and the remote control unit itself. A power management system to conserve unit battery power is also responsive to the pattern recognizer to allow intelligent power management control. The control system uses the display of the consumer electronic equipment to provide instructions to the user, and the behavior of the remote control system uses what is displayed on the display as context information for pattern recognition.

Abstract: A remote control unit selectively transmits a control signal for remotely controlling an electronic device. The unit defines an imaginary cut plane that substantially bisects the unit. The unit includes a plurality of input features collectively disposed symmetrically with respect to the imaginary cut plane. The input features include a first and second input feature. The first and second input features are disposed on opposite sides of the cut plane. Furthermore, the unit includes a sensor that detects a first and second holding position of the unit. The first holding position and the second holding position are substantially opposite to each other. Moreover, the unit includes a controller that associates the control signal with the first input feature when the sensor detects the first holding position, and the controller associates the control signal with the second input feature when the sensor detects the second holding position.

Abstract: Accordingly, a virtual keypad system for inputting text is provided. A virtual keypad system includes a remote controller having at least one touchpad incorporated therein and divided into a plurality of touch zones. A display device is in data communication with the remote controller and is operable to display a user interface including a keypad, where each key of the keypad is mapped to a touch zone of the touchpad. A prediction module, in response to an operator pressing a given touch zone to select a particular character, performs one or more key prediction methods to predict one or more next plausible keys. A key mapping module remaps the touch zones of the touchpad to the keys of the keypad based on the one or more next plausible keys.

Abstract: A language model back-off system can be used with a user interface employing one or more language models to constrain navigation of selectable user interface input components. A user input interpretation module receives user input and interprets the user input to determine if a selection is made of one or more user interface input components. If a selection is not made, the user input interpretation module determines whether conditions are met for backing off one or more language models employed to constrain navigation of the user interface input components. If the conditions are met, a language model back-off module backs off the one or more language models.

Abstract: A method for discriminatively training acoustic models is provided for automated speaker verification (SV) and speech (or utterance) verification (UV) systems.

Abstract: An improved method is provided for constructing compact acoustic models for use in a speech recognizer. The method includes: partitioning speech data from a plurality of training speakers according to at least one speech related criteria (i.e., vocal tract length); grouping together the partitioned speech data from training speakers having a similar speech characteristic; and training an acoustic bubble model for each group using the speech data within the group.

Abstract: A media capture device has an audio input receptive of user speech relating to a media capture activity in close temporal relation to the media capture activity. A plurality of focused speech recognition lexica respectively relating to media capture activities are stored on the device, and a speech recognizer recognizes the user speech based on a selected one of the focused speech recognition lexica. A media tagger tags captured media with generated speech recognition text, and a media annotator annotates the captured media with a sample of the user speech that is suitable for input to a speech recognizer. Tagging and annotating are based on close temporal relation between receipt of the user speech and capture of the captured media. Annotations may be converted to tags during post processing, employed to edit a lexicon using letter-to-sound rules and spelled word input, or matched directly to speech to retrieve captured media.

Abstract: A noise robustness method operates jointly in a signal domain and a model domain. For example, energy is added in the signal domain for frequency bands where an actual noise level of an incoming signal is lower than a noise level used to train models, thus obtaining a compensated signal. Also, energy is added in the model domain for frequency bands where noise level of the incoming signal or the compensated signal is higher than the noise level used to train the models. Moreover, energy is never removed, thereby avoiding problems of higher sensitivity of energy removal to estimation errors.

Abstract: Model compression is combined with model compensation. Model compression is needed in embedded ASR to reduce the size and the computational complexity of compressed models. Model-compensation is used to adapt in real-time to changing noise environments. The present invention allows for the design of smaller ASR engines (memory consumption reduced to up to one-sixth) with reduced impact on recognition accuracy and/or robustness to noises.

Abstract: A method for discriminatively training acoustic models is provided for automated speaker verification (SV) and speech (or utterance) verification (UV) systems.

Abstract: An improved discriminative training method is provided for hidden Markov models. The method includes: defining a measure of separation margin for the data; identifying a subset of training utterances having utterances misrecognized by the models; defining a training criterion for the models based on maximizing the separation margin; formulating the training criterion as a constrained minimax optimization problem; and solving the constrained minimax optimization problem over the subset of training utterances, thereby discriminatively training the models.

Abstract: A method for performing noise adaptation of a target speech signal input to a speech recognition system, where the target speech signal contains both additive and convolutional noises. The method includes estimating an additive noise bias and a convolutional noise bias; in the target speech signal; and jointly compensating the target speech signal for the additive and convolutional noise biases in a feature domain.

Abstract: The dynamic programming technique employs a lexical tree that is encoded in computer memory as a flat representation in which the nodes of each generation occupy contiguous memory locations. The traversal algorithm employs a set of traversal rules whereby nodes of a given generation are processed before the parent nodes of that generation. The deepest child generation is processed first and traversal among nodes of each generation proceeds in the same topological direction.

Abstract: A method and apparatus for data entry by voice under adverse conditions is disclosed. More specifically it provides a way for efficient and robust form filling by voice. A form can typically contain one or several fields that must be filled in. The user communicates to a speech recognition system and word spotting is performed upon the utterance. The spotted words of an utterance form a phrase that can contain field-specific values and/or commands. Recognized values are echoed back to the speaker via a text-to-speech system. Unreliable or unsafe inputs for which the confidence measure is found to be low (e.g. ill-pronounced speech or noises) are rejected by the spotter. Speaker adaptation is furthermore performed transparently to improve speech recognition accuracy. Other input modalities can be additionally supported (e.g. keyboard and touch-screen). The system maintains a dialogue history to enable editing and correction operations on all active fields.

Abstract: Dynamically constructed grammar-constraints and frequency or statistics-based constraints are used to constrain the speech recognizer and to optionally rescore the output to improve recognition accuracy. The recognition system is well adapted for hands-free operation of portable devices, such as for voice dialing operations.